Explaining why some people hear “Yanni” or “Laurel” is actually quite complicated. It involves physics, audio theory, psychoacoustics, and linguistics, but from where I sit as a sound artist, the primary reason we hear either Yanni or Laurel has to do with what’s known as frequency response, which is the measurement of amplitudes over a frequency spectrum. It’s a term usually associated with microphones and loudspeakers, and is one of the primary specifications that creates an audio device’s uniqueness. A frequency response plot is a graph that shows how an audio device will respond at certain frequencies. For example, some microphones are equally sensitive throughout a spectrum (flat FQ response) while other microphones have a more shaped response, perhaps more sensitive to lower frequencies or higher frequencies. The frequency response is an important factor when an audio engineer chooses one microphone over another for a specific instrument or application.
Our ears are our “microphones”, and as such also have a frequency response. Some people are more sensitive to certain bands of frequencies, or groupings of frequencies, than other people. As we get older, this frequency response changes (remember the ringtone that only kids could hear a few years ago?).
In the Yanni/Laurel audio, there is something about the formant (the most fundamental frequencies of a sound) that obfuscates what some people hear. Interestingly, if one lowers the pitch of the formant frequency, more people will begin to hear “Laurel”, and if you raise the pitch most people will begin to hear “Yanni”. Likewise simply filtering out higher frequencies will result in more people hearing “Laurel”, and conversely, filtering out lower frequencies will result in people hearing “Yanni”. We can conclude that people more sensitive to higher frequencies will hear “Yanni” of the original recording, and those more sensitive to lower frequencies will hear “Laurel”.
The NY Times has a more thorough explanation and a really cool tool on their site that allows you to change the audio in real time: